Reported by others
Abstract grid puzzles (ARC-AGI-2)
The model sees a few example grids of colored cells, works out the transformation rule and applies it to a new grid.
How to read this table
- Higher rows are rated higher. A range like “1–2” means the gap is within the margin of error.
- The score is the value published by the task's authors and other organizations.
- Conditions can differ between models, so small gaps do not mean much.
- Prices are a guide per million tokens, reading then writing.
- Length is how much text the model can read at once (tokens).
| Rank | Model | Developer | Accuracy (%) | Input $ /1M | Output $ /1M | Context | Supports | Weights | Released | Price and specs |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | GPT-6 Astra | OpenAI | Rating 95.0 | $10 | $50 | 1.05M | OpenAI Released 2026-09 Price $10 / $50 Length 1.05M | |||
| 2 | GPT-6.1 Sol | OpenAI | Rating 94.2 | $2 | $10 | 1.05M | OpenAI Released 2026-09 Price $2 / $10 Length 1.05M | |||
| 3 | Claude Opus 5.5 | Anthropic | Rating 93.3 | $4 | $20 | 1M | Anthropic Released 2026-09 Price $4 / $20 Length 1M | |||
| 4 | GPT-5.6 Sol | OpenAI | Rating 92.5 | $4 | $20 | 1.05M | OpenAI Released 2026-07 Price $4 / $20 Length 1.05M | |||
| 5 | Gemini 4 Argon | Rating 91.3 | $2 | $10 | 1.05M | Google Released 2026-09 Price $2 / $10 Length 1.05M | ||||
| 6 | Claude Opus 5 | Anthropic | Rating 90.4 | $5 | $25 | 1M | Anthropic Released 2026-07 Price $5 / $25 Length 1M | |||
| 7 | Claude Fable 5.1 | Anthropic | Rating 90.0 | $10 | $50 | 1M | Anthropic Released 2026-09 Price $10 / $50 Length 1M | |||
| 8 | GPT-6 Sol | OpenAI | Rating 89.6 | $2 | $10 | 1.05M | OpenAI Released 2026-09 Price $2 / $10 Length 1.05M | |||
| 9–10 | Claude Fable 5 | Anthropic | Rating 89.2 | $10 | $50 | 1M | Anthropic Released 2026-06 Price $10 / $50 Length 1M | |||
| 9–10 | Gemini 3.8 Flash | Rating 89.2 | $0.75 | $3.75 | 1.05M | Google Released 2026-09 Price $0.75 / $3.75 Length 1.05M | ||||
| 11 | GPT-5.5 | OpenAI | Rating 85.0 | $5 | $30 | 1.05M | OpenAI Released 2026-04 Price $5 / $30 Length 1.05M | |||
| 12 | Gemini 3.7 Flash | Rating 84.6 | $0.75 | $3.75 | 1.05M | Google Released 2026-08 Price $0.75 / $3.75 Length 1.05M | ||||
| 13–14 | Gemini 3 Deep Think | Rating 84.6 | — | — | — | Google Released 2026-02 Price — Length — | ||||
| 13–14 | GPT-5.5 Pro | OpenAI | Rating 84.6 | $30 | $180 | 1.05M | OpenAI Released 2026-04 Price $30 / $180 Length 1.05M | |||
| 15 | GPT-5.6 Terra | OpenAI | Rating 83.9 | $2 | $12 | 1.05M | OpenAI Released 2026-07 Price $2 / $12 Length 1.05M | |||
| 16 | GPT-5.4 Pro | OpenAI | Rating 83.3 | $30 | $180 | 1.05M | OpenAI Released 2026-03 Price $30 / $180 Length 1.05M | |||
| 17 | Gemini 3.1 Pro (preview) | Rating 77.1 | $2 | $12 | 1.05M | Google Released 2026-02 Price $2 / $12 Length 1.05M | ||||
| 18 | Claude Opus 4.7 | Anthropic | Rating 75.8 | $5 | $25 | 1M | Anthropic Released 2026-04 Price $5 / $25 Length 1M | |||
| 19 | GPT-5.4 | OpenAI | Rating 74.0 | $2.5 | $15 | 1.05M | OpenAI Released 2026-03 Price $2.5 / $15 Length 1.05M | |||
| 20 | DeepSeek-V4.1 Flash | DeepSeek | Rating 72.9 | — | — | 1M | DeepSeek Released 2026-09 Price — Length 1M | |||
| 21–22 | Claude Opus 4.8 | Anthropic | Rating 72.1 | $5 | $25 | 1M | Anthropic Released 2026-05 Price $5 / $25 Length 1M | |||
| 21–22 | Gemini 3.5 Flash | Rating 72.1 | $1.5 | $9 | 1.05M | Google Released 2026-05 Price $1.5 / $9 Length 1.05M | ||||
| 23 | Claude Opus 4.6 | Anthropic | Rating 69.2 | $5 | $25 | 1M | Anthropic Released 2026-02 Price $5 / $25 Length 1M | |||
| 24 | Grok 4.6 | xAI | Rating 67.1 | $2 | $6 | 500K | xAI Released 2026-08 Price $2 / $6 Length 500K | |||
| 25 | GLM-5.3 Flash | Z.ai | Rating 65.8 | $0.15 | $0.5 | 1M | Z.ai Released 2026-08 Price $0.15 / $0.5 Length 1M | |||
| 26 | Grok 4.20 | xAI | Rating 65.1 | — | — | 2M | xAI Released 2026-02 Price — Length 2M | |||
| 27–28 | DeepSeek-V4 Flash (2026-07) | DeepSeek | Rating 61.4 | — | — | 1M | DeepSeek Released 2026-07 Price — Length 1M | |||
| 27–28 | Grok 4.7 | xAI | Rating 61.4 | $2 | $6 | 500K | xAI Released 2026-09 Price $2 / $6 Length 500K | |||
| 29–30 | Claude Haiku 5.5 | Anthropic | Rating 61.3 | $0.1 | $0.5 | 1M | Anthropic Released 2026-10 Price $0.1 / $0.5 Length 1M | |||
| 29–30 | DeepSeek-V4 Pro (2026-08) | DeepSeek | Rating 61.3 | $0.66 | $1.98 | 1M | DeepSeek Released 2026-08 Price $0.66 / $1.98 Length 1M | |||
| 31 | Claude Sonnet 4.6 | Anthropic | Rating 60.4 | $3 | $15 | 1M | Anthropic Released 2026-02 Price $3 / $15 Length 1M | |||
| 32–33 | Gemini 3.6 Flash | Rating 60.4 | $0.75 | $3.75 | 1.05M | Google Released 2026-07 Price $0.75 / $3.75 Length 1.05M | ||||
| 32–33 | Kimi K3 | Moonshot | Rating 60.4 | $3 | $15 | 1.05M | Moonshot Released 2026-07 Price $3 / $15 Length 1.05M | |||
| 34 | GPT-5.6 Luna | OpenAI | Rating 59.5 | $0.2 | $1.2 | 1.05M | OpenAI Released 2026-07 Price $0.2 / $1.2 Length 1.05M | |||
| 35 | GPT-6 Luna | OpenAI | Rating 59.3 | $0.1 | $0.5 | 1.05M | OpenAI Released 2026-09 Price $0.1 / $0.5 Length 1.05M | |||
| 36 | GPT-5.2 Pro | OpenAI | Rating 54.2 | $21 | $168 | 400K | OpenAI Released 2025-12 Price $21 / $168 Length 400K | |||
| 37 | GPT-5.2 | OpenAI | Rating 52.9 | $1.75 | $14 | 400K | OpenAI Released 2025-12 Price $1.75 / $14 Length 400K | |||
| 38 | Grok 4.5 | xAI | Rating 52.6 | $2 | $6 | 500K | xAI Released 2026-07 Price $2 / $6 Length 500K | |||
| 39 | Qwen3.8 27B | Alibaba | Rating 42.4 | — | — | 262.1K | Alibaba Released 2026-08 Price — Length 262.1K | |||
| 40 | Inkling Small | Thinking Machines | Rating 40.1 | $0.58 | $1.44 | 524.3K | Thinking Machines Released 2026-07 Price $0.58 / $1.44 Length 524.3K | |||
| 41 | Claude Opus 4.5 | Anthropic | Rating 37.6 | $5 | $25 | 200K | Anthropic Released 2025-11 Price $5 / $25 Length 200K | |||
| 42 | Inkling | Thinking Machines | Rating 36.5 | $1.87 | $4.68 | 262.1K | Thinking Machines Released 2026-07 Price $1.87 / $4.68 Length 262.1K | |||
| 43 | Gemini 3 Flash (preview) | Rating 33.6 | $0.5 | $3 | 1M | Google Released 2025-12 Price $0.5 / $3 Length 1M | ||||
| 44 | Gemini 3 Pro (preview) | Rating 31.1 | — | — | 1M | Google Released 2025-11 Price — Length 1M | ||||
| 45 | GLM-5.2 | Z.ai | Rating 22.8 | $1.4 | $4.4 | 1M | Z.ai Released 2026-06 Price $1.4 / $4.4 Length 1M | |||
| 46 | GPT-5.4 mini | OpenAI | Rating 18.9 | $0.75 | $4.5 | 400K | OpenAI Released 2026-03 Price $0.75 / $4.5 Length 400K | |||
| 47 | GPT-5 Pro | OpenAI | Rating 18.3 | $15 | $120 | 400K | OpenAI Released 2025-10 Price $15 / $120 Length 400K | |||
| 48 | GPT-5.1 | OpenAI | Rating 17.6 | $1.25 | $10 | 400K | OpenAI Released 2025-11 Price $1.25 / $10 Length 400K | |||
| 49 | Grok 4 | xAI | Rating 16.0 | — | — | 256K | xAI Released 2025-07 Price — Length 256K | |||
| 50 | Claude Sonnet 4.5 | Anthropic | Rating 13.6 | $3 | $15 | 200K | Anthropic Released 2025-09 Price $3 / $15 Length 200K | |||
| 51 | Kimi K2.5 | Moonshot | Rating 11.8 | — | — | 262.1K | Moonshot Released 2026-01 Price — Length 262.1K | |||
| 52 | Gemini 3.5 Flash-Lite | Rating 10.3 | $0.3 | $2.5 | 1.05M | Google Released 2026-07 Price $0.3 / $2.5 Length 1.05M | ||||
| 53 | GPT-5 | OpenAI | Rating 9.9 | $1.25 | $10 | 400K | OpenAI Released 2025-08 Price $1.25 / $10 Length 400K | |||
| 54 | Claude Opus 4 | Anthropic | Rating 8.6 | — | — | 200K | Anthropic Released 2025-05 Price — Length 200K | |||
| 55 | o3 | OpenAI | Rating 6.5 | $2 | $8 | 200K | OpenAI Released 2025-04 Price $2 / $8 Length 200K | |||
| 56 | Tiny Recursion Model | Samsung SAIL | Rating 6.3 | — | — | — | — | Samsung SAIL Released 2025-10 Price — Length — | ||
| 57 | o4-mini | OpenAI | Rating 6.1 | $1.1 | $4.4 | 200K | OpenAI Released 2025-04 Price $1.1 / $4.4 Length 200K | |||
| 58 | Claude Sonnet 4 | Anthropic | Rating 5.9 | — | — | 200K | Anthropic Released 2025-05 Price — Length 200K | |||
| 59 | GPT-5.4 nano | OpenAI | Rating 5.7 | $0.2 | $1.25 | 400K | OpenAI Released 2026-03 Price $0.2 / $1.25 Length 400K | |||
| 60 | Grok 4 Fast | xAI | Rating 5.3 | — | — | 2M | xAI Released 2025-09 Price — Length 2M | |||
| 61–64 | Gemini 2.5 Pro | Rating 4.9 | $1.25 | $10 | 1M | Google Released 2025-06 Price $1.25 / $10 Length 1M | ||||
| 61–64 | GLM-5 | Z.ai | Rating 4.9 | $1 | $3.2 | 204.8K | Z.ai Released 2026-02 Price $1 / $3.2 Length 204.8K | |||
| 61–64 | MiniMax-M2.5 | MiniMax | Rating 4.9 | $0.3 | $1.2 | 204.8K | MiniMax Released 2026-02 Price $0.3 / $1.2 Length 204.8K | |||
| 61–64 | o3-pro | OpenAI | Rating 4.9 | — | — | 200K | OpenAI Released 2025-06 Price — Length 200K | |||
| 65 | GPT-5 mini | OpenAI | Rating 4.4 | $0.25 | $2 | 400K | OpenAI Released 2025-08 Price $0.25 / $2 Length 400K | |||
| 66–67 | Claude Haiku 4.5 | Anthropic | Rating 4.0 | $1 | $5 | 200K | Anthropic Released 2025-10 Price $1 / $5 Length 200K | |||
| 66–67 | DeepSeek-V3.2 | DeepSeek | Rating 4.0 | — | — | 128K | DeepSeek Released 2025-12 Price — Length 128K | |||
| 68 | o3-mini | OpenAI | Rating 3.0 | $1.1 | $4.4 | 200K | OpenAI Released 2025-01 Price $1.1 / $4.4 Length 200K | |||
| 69 | GPT-5 nano | OpenAI | Rating 2.6 | $0.05 | $0.4 | 400K | OpenAI Released 2025-08 Price $0.05 / $0.4 Length 400K | |||
| 70 | Gemini 2.5 Flash (preview, 2025-05) | Rating 2.5 | — | — | 1.05M | Google Released 2025-05 Price — Length 1.05M | ||||
| 71–72 | DeepSeek-R1 | DeepSeek | Rating 1.3 | — | — | 128K | DeepSeek Released 2025-01 Price — Length 128K | |||
| 71–72 | Gemini 2.0 Flash | Rating 1.3 | — | — | — | Google Released 2025-02 Price — Length — | ||||
| 73 | Codex Mini | OpenAI | Rating 1.3 | — | — | — | — | OpenAI Released 2025-05 Price — Length — | ||
| 74 | Qwen3 235B A22B (instruct) | Alibaba | Rating 1.3 | — | — | 128K | Alibaba Released 2025-07 Price — Length 128K | |||
| 75 | DeepSeek-R1 (2025-05) | DeepSeek | Rating 1.1 | — | — | 131.1K | DeepSeek Released 2025-05 Price — Length 131.1K | |||
| 76 | Claude 3.7 Sonnet | Anthropic | Rating 0.9 | — | — | 200K | Anthropic Released 2025-02 Price — Length 200K | |||
| 77 | o1-mini | OpenAI | Rating 0.8 | — | — | 128K | OpenAI Released 2024-09 Price — Length 128K | |||
| 78–79 | Gemini 1.5 Pro 002 | Rating 0.8 | — | — | — | Google Released 2024-09 Price — Length — | ||||
| 78–79 | GPT-4.5 (preview) | OpenAI | Rating 0.8 | — | — | — | OpenAI Released 2025-02 Price — Length — | |||
| 80–81 | GPT-4.1 | OpenAI | Rating 0.4 | $2 | $8 | 1.05M | OpenAI Released 2025-04 Price $2 / $8 Length 1.05M | |||
| 80–81 | Grok 3 Mini | xAI | Rating 0.4 | — | — | 131.1K | xAI Released 2025-06 Price — Length 131.1K | |||
| 82–91 | Gemini 2.5 Pro (preview, 2025-05) | Rating 0.0 | — | — | 1.05M | Google Released 2025-05 Price — Length 1.05M | ||||
| 82–91 | GPT-4.1 mini | OpenAI | Rating 0.0 | — | — | 1.05M | OpenAI Released 2025-04 Price — Length 1.05M | |||
| 82–91 | GPT-4.1 nano | OpenAI | Rating 0.0 | $0.1 | $0.4 | 1.05M | OpenAI Released 2025-04 Price $0.1 / $0.4 Length 1.05M | |||
| 82–91 | GPT-4o (2024-11) | OpenAI | Rating 0.0 | $2.5 | $10 | 128K | OpenAI Released 2024-11 Price $2.5 / $10 Length 128K | |||
| 82–91 | GPT-4o mini | OpenAI | Rating 0.0 | — | — | 128K | OpenAI Released 2024-07 Price — Length 128K | |||
| 82–91 | Grok 3 | xAI | Rating 0.0 | — | — | 131.1K | xAI Released 2025-04 Price — Length 131.1K | |||
| 82–91 | Llama 4 Maverick | Meta | Rating 0.0 | — | — | 524.3K | Meta Released 2025-04 Price — Length 524.3K | |||
| 82–91 | Llama 4 Scout | Meta | Rating 0.0 | — | — | 128K | Meta Released 2025-04 Price — Length 128K | |||
| 82–91 | Magistral Medium | Mistral | Rating 0.0 | — | — | — | Mistral Released 2025-06 Price — Length — | |||
| 82–91 | Magistral Small | Mistral | Rating 0.0 | — | — | 128K | Mistral Released 2025-06 Price — Length 128K |
Best accuracy over time
2024-07-18Release date2026-10-07
How to read this page
- The puzzles are easy for people and are built so that memorized knowledge does not solve them. Figures published by the organizations that made the tasks, among others, collected here. The value is accuracy.
- The accuracy is the share of problems solved. A gap of 1 point is 1 question out of 100.
- In the Supports column the icons are, from left: can use tools / can read images / can answer in a fixed format (JSON); faded means not supported or unknown.
- Open-weights models have their contents (the weights) published, so you can also run them on your own computer or server. Models without the open-weights mark are normally used through the developer's own services (apps or API); for some models the weights status is unknown.
- Names such as ChatGPT next to a model are the consumer services its developer offers. Check on the service itself whether that model is available there.
Data: Arena, Epoch AI, Swallow LLM team (CC BY 4.0), Models.dev (MIT), updated 2026-10-24, data sources